A novel algorithm for the authentication of individuals through retinal vascular pattern recognition

This paper describes an original numerical procedure for the recognition of images applied to the case of individuals’ authentication by using biometric patterns, like the retinal vascular one. Because of the statistical nature of the recognition procedure, the 2x2 confusion matrix, typical of decision tests for binary lassification models, is here considered in order to evaluate the performances of the proposed matching algorithm. An ad hoc procedure is then performed both to build the templates of reference database and to evaluate the threshold for the final decision test.